《1 Introduction》

1 Introduction

In recent years, China’s economy has rapidly entered a high-quality development phase with the accelerated growth of Internet, big data, cloud computing, 5G (fifth-generation cellular mobile communications), and other technologies, along with breakthroughs in new-generation AI technologies such as big data-based intelligence, man–machine hybrid intelligence, swarm intelligence, cross-media reasoning, and autonomous intelligence [1]. The deep fusion of manufacturing technology with new-generation intelligent technology and information communication technology is enabling a transformation in the manufacturing development philosophies, manufacturing models, manufacturing approaches, and its ecosystems and is promoting innovative, coordinated, eco-friendly, open, and shared development of the national economy, national welfare and people’s livelihood, and national security.

The new information environment with major changes (such as ubiquitous networks and emerging big data), new technologies (such as improvements in high-performance computing capability and breakthroughs in AI models and algorithms), and new goals of social development (such as intelligent national economy, intelligent national welfare and people’s livelihood, and intelligent defense) are accelerating the development of intelligent manufacturing technology and applications in a brand new era of “new Internet + big data + AI+.” With a strong demand for economic development, the development of a new, intelligent, and interconnected manufacturing system based on human/machine/ material/environment/information for “new Internet + new intelligent manufacturing resources/capabilities/products” becomes the main approach to transform and upgrade the manufacturing industry.

《2 Connotation of intelligent manufacturing system led by new-generation AI technology》

2 Connotation of intelligent manufacturing system led by new-generation AI technology

The following aspects are considered in the connotation of an intelligent manufacturing system guided by new-generation AI technology [2,3]:

(1) Technical Approach: Construction of a user-centered cloud service pool (interconnected service network) of unified operating intelligent manufacturing resources, products, and capabilities where users can obtain intelligent manufacturing services anytime and anywhere on demand through terminals and an intelligent cloud manufacturing service platform. This digital, networked, and intelligent approach is based on ubiquitous networks and represents a deep fusion of new manufacturing technology, new information science & technology, new intelligent science & technology, and manufacturing application-related domain technology.

(2) Characteristics: to conduct intelligent sensing, inter-connecting, cooperating, learning, analyzing, cognizing, decision-making, controlling, and execution on human/machines/ material/environment data in the lifecycle activities (industrial chain) of manufacturing.

(3) Implementation Contents: The integration and optimization of human, technology/equipment, management, data, material, and financial factors (six factors) together with human, technology, management, data, logistics, and financial flows (six flows) in the lifecycle activities of manufacturing.

(4) Paradigm: user-centered, human/machine/material/environment/information-integrated, inter-connected (collaborative), service-oriented, individualized (customized), flexible, socialized, and intelligent manufacturing paradigm.

(5) Ecosystem: Ubiquitous interconnection, data-driven, shared services, cross-field integration, autonomous intelligence, and mass innovation.

(6)Target: To enable efficient, high-quality, low-cost, green, and flexible product manufacturing and user services to improve the market competitiveness of enterprises (or groups).

《3 Architecture of the intelligent manufacturing system led by new-generation AI technology》

3 Architecture of the intelligent manufacturing system led by new-generation AI technology

The architecture of an intelligent manufacturing system led by new-generation AI technology is created according to its connotation. Fig. 1 illustrates the architecture of the intelligent manufacturing system led by new-generation AI technology [2,4], which comprises five layers: new resources/capabilities/products layer, new sensing/access/communications layer, new platform layer, new intelligent-manufacturing cloud services application layer, and new man/organization layer.

《Fig. 1》

Fig. 1. System architecture of an intelligent manufacturing system led by new-generation AI technology.

IaaS: Infrastructure as a Service; DaaS: Data as a Service; PaaS: Platform as a Service; SaaS: Software as a Service; CaaS: Capability as a Service; COaaS: Cooperation as a Service; RFID: Radio Frequency Identification Devices; GPS: Global Positioning System.

《4 Technology system of the intelligent manufacturing system led by new-generation AI technology》

4 Technology system of the intelligent manufacturing system led by new-generation AI technology

Fig. 2 illustrates the technology system of the intelligent manufacturing system led by new-generation AI technology. It primarily comprises the overall technologies, platform technologies, and intelligent technologies for the life cycle activities of manufacturing [2,3], among others.

《Fig. 2》

Fig. 2. Technology system of the intelligent manufacturing system led by new-generation AI technology.

CAX: CAX is the collective name of CAD, CAM, CAE, CAPP, CIM, CIMS, CAS, CAT, CAI, etc., which start with CA, while X represents all; DFX: DFX is an abbreviation for “Design for X” (Design for each/a segment of the product life cycle), in which X may represent the life cycle of a product or one of the links, such as mounting (M-manufacturing, T-testing), processing, use, maintenance, recovery and rejection, and may also represent the product competitiveness or factors that determine the competitive strength, such as quality, cost and time, of a product.

《5 User case for the intelligent manufacturing system led by new-generation AI technology》

5 User case for the intelligent manufacturing system led by new-generation AI technology

Based on our research on the China Aerospace Science and Industry Corporation (CASIC) in the industrial Internet, we integrate the intelligent manufacturing system led by new-generation AI technology into the development of the first Chinese industrial Internet system: CASICloud [5,6], which is an industrial Internet designed for Chinese industrial development. As for the weak foundation of industrial development in China and the co-existence of industrial development 1.0/2.0/3.0, CASICloud is working on parallel development of digital, networked, and intelligent manufacturing to boost the all-round transformation and upgradation of the manufacturing industry. This paper focuses on the architecture of CASICloud, the industrial big data/AI subsystem developed on the CASICloud platform, and the application cases and achievements of the big data-based intelligence technology.

《5.1 CASICloud system architecture》

5.1 CASICloud system architecture

Fig. 3 illustrates the architecture of the CASICloud system.

《Fig. 3》

Fig. 3. CASICloud architecture.

Native Apps: This is a third-party application based on the smartphone’s native operating system such as iOS, Android, or WP and runs with native programming; CAD: Computer Aided Design; CAE: Computer Aided Engineering; CAPP: Computer Aided Process Planning; CAM: Computer Aided Manufacturing; SCM/SRM: Supply Chain Management/Supplier Relationship Management; B2B2C: Business to Business to Customer; ERP: Enterprise Resource Planning; CRP: Cloud Resource Planning; MES: Manufacturing Execution System; C-CRM: Cloud-Customer Relationship Management; CIS: Corporate Identity System; Open 3rd Apps: Open third-party Apps; AOP: Astronautic Open Platform; Cloud Foundry: Open source PaaS cloud platform from VMware; HDFS: Hadoop Distributed File System; MongoDB: a database based on distributed file storage, written with C ++ language, with the aim of providing a scalable and high-performance data storage solution for web applications; Cassandra: An open-source NoSQL distributed database system; HBase: a distributed, column-oriented open-source database; MapReduce: a programming model for parallel computing of large datasets (over 1 TB). Spak: a fast, universal computing engine designed for large-scale data processing; Storm: A distributed real-time big data processing system; Hive: a data warehouse tool based on Hadoop; Mlib: a machine learning library; Ambari: A web-based tool supporting the supply, management, and monitoring of Apache Hadoop cluster; CDH: Cloudera’s Distribution Including Apache Hadoop; INDICS-APIs: Industrial Intelligent Cloud System Platform−Application Programming Interfaces; SCADA: Supervisory Control and Data Acquisition; OPC UA: OPC Unified Architecture; Profinet: new-generation automation bus standard developed by PROFIBUS International (PI) based on industrial Ethernet technology and an on-board terminal extension protocol; MQTT: Message Queuing Telemetry Transport; JTEXT: a vehicle terminal extension protocol; Modbus: invented by Modicon (now a brand of Schneider Electric), bus protocol for industrial sites; NB-IOT: Narrow Band Internet of Things; EDP: Embedded Display Port; HTTP: Hyper Text Transfer Protocol; DTU: Data Transfer unit; WIA: Wireless Networks for Industrial Automation; C-Resources: CloudResources; TCP/IP: Transmission Control Protocol/Internet Protocol; Profibus: a fieldbus standard for automation technology.

《5.2 Intelligent manufacturing application case based on big data intelligence technology》

5.2 Intelligent manufacturing application case based on big data intelligence technology

5.2.1 Industrial big data/AI subsystem of CASICloud

Fig. 4 presents the architecture of industrial big data/AI subsystem based on the application of big data intelligence technology on the CASICloud platform.

《Fig. 4》

Fig. 4. Industry big data/AI subsystem for CASICloud.

OpenAPI: Open Application Programming Interfaces; BDE: Borland Database Engine; MySQL: a relational database management system; PostgreSQL: An open-source object relational database system; HTTPS: Hyper Text Transfer Protocol over Secure Socket Layer; MQTT: Message Queuing Telemetry Transport; AMQP: Advanced Message Queuing Protocol; IT: Information Technology.

5.2.2 Key big data intelligence technologies for CASICloud

With emphasis on the requirements of industrial big data/AI subsystem of CASICloud, we study big data acquisition, storage, management, analysis, application technology, and key technologies to ensure adherence to big data standards, quality, and security.

In this research on big data integration and cleaning, we use Extract-Transform-Load (ETL) to clean and convert redundant and dirty data created upon operating the equipment. We monitor data quality for big enterprise’s data storage and management using time series, columnar database, distributed historical data storage, and document database. We implement offline analysis and mining for overall equipment effectiveness (OEE) using Hadoop. We implement real-time fault diagnosis using Storm. We visualize enterprise data using business intelligence (BI). We study data encryption, backup, and access authority for guaranteeing data security. Using these big data intelligence technologies, CASICloud can optimize, upgrade, and enhance the competitiveness of enterprises through the intelligent application of enterprise informatization, industrial IoT (Internet of Things) equipment and cross-border data.

5.2.3 Applications

Currently, the industrial subsystem of CASICloud has achieved certain results in research and development (R&D), production, operation, and maintenance. For example, during an aerospace cable design, we saved over 60% of the development time using the cable design tool E3 based on historical big data and integrated flight performance data. With adoption of the CASICloud-based industrial big data/AI subsystem, cloud scheduling of aerospace manufacturing is realized. According to the multi-unit large-scale coordination demand of aerospace products, a cloud-based manufacturing execution system is adopted to collect and monitor data from production and processing centers of multiple enterprises, and collaborative and optimization scheduling of production resources across different enterprises is realized based on a cloud scheduling algorithm.

《6 Suggestions for the development of an intelligent manufacturing system led by newgeneration AI technology》

6 Suggestions for the development of an intelligent manufacturing system led by newgeneration AI technology

After research and preliminary practice, we conclude that the intelligent manufacturing system led by new-generation AI technology is a new technical route to accelerate the development of intelligent manufacturing. This embodies the characteristics of the “new Internet + big data + AI+” era in terms of mode, means, and format.

With the fast pace of innovation-driven development and deep integration of industrialization and informatization, persistence is required in the integrated and coordinated development of technology, applications, and industry to promote sustainable and steady development of an intelligent manufacturing system led by new-generation AI technology. Additionally, in China’s market-based open innovation system, compatibility between the characteristics of countries, industries, and enterprises is required to promote the development and implementation of an intelligent manufacturing system.